no code implementations • 5 Mar 2024 • Mrinmay Sen, A. K. Qin, Gayathri C, Raghu Kishore N, Yen-Wei Chen, Balasubramanian Raman
This paper introduces a new stochastic optimization method based on the regularized Fisher information matrix (FIM), named SOFIM, which can efficiently utilize the FIM to approximate the Hessian matrix for finding Newton's gradient update in large-scale stochastic optimization of machine learning models.
1 code implementation • 12 Feb 2024 • Puneet Kumar, Sarthak Malik, Balasubramanian Raman, Xiaobai Li
It implements an interpretability technique to analyze the contribution of textual and visual features during the generation of uncontrolled and controlled feedback.
no code implementations • 1 Jul 2023 • Kishore Babu Nampalle, Pradeep Singh, Uppala Vivek Narayan, Balasubramanian Raman
In the rapidly evolving landscape of medical imaging diagnostics, achieving high accuracy while preserving computational efficiency remains a formidable challenge.
no code implementations • 30 Jun 2023 • Kishore Babu Nampalle, Pradeep Singh, Uppala Vivek Narayan, Balasubramanian Raman
The proliferation of deep learning applications in healthcare calls for data aggregation across various institutions, a practice often associated with significant privacy concerns.
no code implementations • 27 Jun 2023 • Pradeep Singh, Kishore Babu Nampalle, Uppala Vivek Narayan, Balasubramanian Raman
In this paper, we propose a novel curriculum learning-based approach to train deep learning models to handle occluded medical images effectively.
no code implementations • 17 May 2023 • Kishore Babu Nampalle, Pradeep Singh, Vivek Narayan Uppala, Sumit Gangwar, Rajesh Singh Negi, Balasubramanian Raman
In healthcare, accurately classifying medical images is vital, but conventional methods often hinge on medical data with a consistent grid structure, which may restrict their overall performance.
1 code implementation • 25 Aug 2022 • Puneet Kumar, Sarthak Malik, Balasubramanian Raman
A new interpretability technique has been developed to identify the important speech & image features leading to the prediction of particular emotion classes.
1 code implementation • 24 Aug 2022 • Puneet Kumar, Sarthak Malik, Balasubramanian Raman, Xiaobai Li
This paper proposes a multimodal emotion recognition system, VIsual Spoken Textual Additive Net (VISTA Net), to classify emotions reflected by multimodal input containing image, speech, and text into discrete classes.
1 code implementation • 23 Mar 2022 • Puneet Kumar, Gaurav Bhat, Omkar Ingle, Daksh Goyal, Balasubramanian Raman
A feedback synthesis system has been proposed and trained using ground-truth human comments along with image-text input.
no code implementations • 23 Mar 2021 • Anshul Pundhir, Deepak Verma, Puneet Kumar, Balasubramanian Raman
This paper has proposed a novel approach to classify the subjects' smoking behavior by extracting relevant regions from a given image using deep learning.
no code implementations • 17 Nov 2020 • Puneet Kumar, Balasubramanian Raman
It addresses the challenge of the insufficient availability of pre-trained models and well-annotated datasets for image emotion recognition (IER).
no code implementations • 6 Sep 2020 • Vipul Bansal, Himanshu Buckchash, Balasubramanian Raman
Evaluation of students' performance for the completion of courses has been a major problem for both students and faculties during the work-from-home period in this COVID pandemic situation.
no code implementations • 21 Sep 2019 • Vivekraj V. K., Debashis Sen, Balasubramanian Raman
Video skimming, also known as dynamic video summarization, generates a temporally abridged version of a given video.
no code implementations • 21 Jan 2018 • Gaurav Bhatt, Shivam Sharma, Balasubramanian Raman
Further, we use the tensor parameters to introduce a 3-way interaction between question, answer and external features in vector space.
1 code implementation • 11 Dec 2017 • Gaurav Bhatt, Aman Sharma, Shivam Sharma, Ankush Nagpal, Balasubramanian Raman, Ankush Mittal
We present a novel idea that combines the neural, statistical and external features to provide an efficient solution to this problem.
Ranked #3 on Fake News Detection on FNC-1
1 code implementation • 31 Oct 2017 • Gaurav Bhatt, Piyush Jha, Balasubramanian Raman
In a broader perspective, the techniques used to investigate common representation learning falls under the categories of canonical correlation-based approaches and autoencoder based approaches.